Annual and Seasonal Extreme Precipitation Probability Distributions at Thessaloniki Based Upon Hourly Values

The objective of this paper is to identify the best fit probability distribution of extreme precipitation conditions in Thessaloniki, for both annual and seasonal basis, by using precipitation measurements obtained at the Aristotle University of Thessaloniki meteorological station, during the period 1947–2003. A new threshold determining method is proposed, in order to define the extreme precipitation probability distributions in Thessaloniki, based upon the polynomial trends of the cumulative distribution functions, retrieved from hourly precipitation measurements. Through this approach, three statistical goodness of fit tests were carried out, in order to come up with the best fitting probability distribution out of more than 30 probability distributions, for annual and seasonal extreme precipitation conditions. The best fit probability distribution for each study period was determined, through the calculation of the minimum absolute deviation between actual and estimated values, being produced by generating random numbers from the parameters of the selected probability distributions. Based upon this fitting distribution procedure, return periods for different precipitation extreme values were calculated. Results reveal the Generalized Pareto distribution as the best fit probability model for the annual extreme precipitation, and the Johnson SB, log-Gamma and Lognormal distributions for winter and spring, summer, and autumn, respectively.